Estrogen, synaptic plasticity and hypothalamic reproductive aging.

Unlike primates who undergo ovarian failure and loss of sex steroids at the end of reproduction, aging rodents undergo constant vaginal estrus followed by constant diestrus and finally anestrus, which indicates the absence of responsive ovarian follicles. The latter state is analogous to menopause in women. The timing of the appearance of constant estrus is determined by many factors including estrogen exposure in the brain during development and the number of times that the animal gets pregnant. The chief site of this reproductive aging in rat brains is the arcuate nucleus of the hypothalamus. The transition from normal cycles to constant estrus parallels the females' gradually decreased ability to respond to administered estradiol with a cycle of inhibition followed by disinhibition of gonadotrophin-releasing hormone. Evidence has accumulated indicating this to be due to a loss of the rat's ability to respond to markedly elevated estradiol with the usual arcuate nucleus neuro-glial plasticity that supports the estrogen-induced gonadotrophin surge (EIGS). Just as male rats are not capable of an EIGS, aged females loose this ability through repeated EIGS. Experiments indicate that in male rats the hypothalamic synaptology that develops as a result of exposure to testicular androgens in the perinatal period (brain sexual differentiation) is a result of conversion of testosterone from the testes to estrogen in the brain and is therefore due to early estrogen exposure. Aging females appear to reach a synaptology similar to males and constant estrus as a result of repeated exposure to ovarian estrogens during their reproductive careers. The relative role of aging and hormonal factors remains unclear. Morphological evidence is presented that indicates the above effects of estrogen involve changes in hypothalamic arcuate nucleus neurons and glia, including changes in the organization of perikaryal membranes as well as arcuate nucleus synaptology and the load of peroxidase in the astroglia. A possible role for free radicals (reactive oxygen species) in hypothalamic reproductive aging has been proposed. Such a mechanism is supported by evidence that the anti-oxidant vitamin E delays the onset of constant estrus and the accumulation of glial peroxidase in aging female rats. However, since the synaptology and peroxidase load in constant estrus females is independent of the age at which the constant estrus occurs, it appears that the role of (repeated) estradiol exposure is more deterministic of hypothalamic failure than is aging, per se.

Hormonal regulation of hippocampal spine synapse density involves subcortical mediation.

It is well established that estrogen has positive effects on the density of pyramidal cell spines in the hippocampal CA1 subfield. This study explored whether afferent connections of the hippocampus that come from estrogen-sensitive subcortical structures, including the septal complex, median raphe and supramammillary area, play a role in this estrogen-induced hippocampal synaptic plasticity. These particular subcortical structures have major influences on hippocampal activity, including theta rhythm and long-term potentiation. The latter also promotes the formation of new synapses. All of the rats were ovariectomized; the fimbria/fornix, which contains the majority of subcortical efferents to the hippocampus, was transected unilaterally in each, and half of the animals received estrogen replacement. Using unbiased electron microscopic stereological methods, the CA1 pyramidal cell spine synapse density was calculated. In the estrogen-treated rats, contralateral to the fimbria/fornix transection, the spine density of CA1 pyramidal cells increased dramatically, compared to the spine density values of both the ipsilateral and contralateral hippocampi of non-estrogen-treated animals and to that of the ipsilateral hippocampus of the estrogen replaced rats. These observations indicate that fimbria/fornix transection itself does not considerably influence CA1 area pyramidal cell spine density and, most importantly, that the estrogenic effect on hippocampal morphology, in addition to directly affecting the hippocampus, involves subcortical mediation.

The supramammillo-hippocampal and supramammillo-septal glutamatergic/aspartatergic projections in the rat: a combined [3H]D-aspartate autoradiographic and immunohistochemical study.

It is well established that the supramammillary nucleus plays a critical role in hippocampal theta rhythm generation/regulation by its direct and indirect (via the septal complex) connections to the hippocampus. Previous morphological and electrophysiological studies indicate that both the supramammillo-hippocampal and supramammillo-septal efferents contain excitatory transmitter. To test the validity of this assumption, transmitter specific retrograde tracer experiments were performed. [3H]D-aspartate was injected into different locations of the hippocampus (granular and supragranular layers of the dentate gyrus and CA2 and CA3a areas of the Ammon's horn) and septal complex (medial septum and the area between the medial and lateral septum) that are known targets of the supramammillary projection. Consecutive vibratome sections prepared from the entire length of the posterior hypothalamus, including the supramammillary area, were immunostained for calretinin, tyrosine hydroxylase, or calbindin, and further processed for autoradiography. Radiolabeled, radiolabeled plus calretinin-containing, and calretinin-immunoreactive neurons were plotted at six different oro-caudal levels of the supramammillary area. The results demonstrated that following both hippocampal and septal injection of the tracer, the majority of the retrogradely radiolabeled (glutamatergic/aspartatergic) cells are immunoreactive for calretinin. However, non-radiolabeled calretinin-containing neurons and radiolabeled calretinin-immunonegative cells were also seen, albeit at a much lower density. These observations clearly indicate the presence of glutamatergic/aspartatergic projections to both the hippocampus and septal complex. It may be assumed that this transmitter could play a role in hippocampal theta rhythm generation/regulation.

Opioids suppress IPSCs in neurons of the rat medial septum/diagonal band of Broca: involvement of mu-opioid receptors and septohippocampal GABAergic neurons.

The medial septum/diagonal band region (MSDB), which provides a major cholinergic and GABAergic input to the hippocampus, expresses a high density of opioid receptors. Behaviorally, intraseptal injections of opioids produce deficits in spatial memory, however, little is known about the electrophysiological effects of opioids on MSDB neurons. Therefore, we investigated the electrophysiological effects of opioids on neurons of the MSDB using rat brain slices. In voltage-clamp recordings with patch electrodes, bath-applied met-enkephalin, a nonselective opioid receptor agonist, decreased the number of tetrodotoxin and bicuculline-sensitive inhibitory synaptic currents in cholinergic- and GABA-type MSDB neurons. A similar effect occurred in brain slices containing only the MSDB, suggesting that opioids decrease GABA release primarily by inhibiting spontaneously firing GABAergic neurons located within the MSDB. Accordingly, in extracellular recordings, opioid-sensitive, spontaneously firing neurons could be found within the MSDB. Additionally, in intracellular recordings a subpopulation of GABA-type neurons were directly inhibited by opioids. All effects of met-enkephalin were mimicked by a mu receptor agonist, but not by delta or kappa agonists. In antidromic activation studies, mu-opioids inhibited a subpopulation of septohippocampal neurons with high conduction velocity fibers, suggestive of thickly myelinated GABAergic fibers. Consistent with the electrophysiological findings, in double-immunolabeling studies, 20% of parvalbumin-containing septohippocampal GABA neurons colocalized the mu receptor, which at the ultrastructural level, was found to be associated with the neuronal cell membrane. Thus, opioids, via mu receptors, inhibit a subpopulation of MSDB GABAergic neurons that not only make local connections with both cholinergic and noncholinergic-type MSDB neurons, but also project to the hippocampus.

Extrinsic and intrinsic substance P innervation of the rat lateral septal area calbindin cells.

The electrophysiological observations that substance P administration to the lateral septal area elicits both excitatory and inhibitory responses, together with earlier reports on the multiple sources of substance P innervation of the septum, implies that these axons with distinct origins have different functions. This prompted us to examine the origin and neurochemical character of substance P afferents to the lateral septal area. Chronic surgical isolation of the septum from its ventral afferents and retrograde tracer experiments using wheat germ agglutinin-conjugated horseradish peroxidase, both followed by an immunostaining for substance P, were employed to elucidate the origin of these axon terminals. In order to assess the possible co-existence of substance P with other neurotransmitter substances in the parent cells of the septopetal projections, co-localization studies for substance P and choline acetyltransferase, as well as substance P and GABA, were performed. The comparative distribution of substance P fibers and septal calbindin-containing neurons was also investigated using correlated light and electron microscopic double immunostaining. The results are summarized as follows: (i) the substance P innervation of the lateral septal area derives from several hypothalamic nuclei (including the lateral and lateroanterior hypothalamic area, tuber cinereum and ventromedial hypothalamic nucleus) and tegmental nuclei (the majority of fibers from the laterodorsal and a few from the pedunculopontine tegmental nucleus), as well as intrinsic septal cells; (ii) the septopetal substance P fibers of tegmental origin are cholinergic; intraseptal substance P neurons located in the dorsolateral part of the lateral septum also contain GABA, while substance P neurons seen on the border between the medial and lateral septal area and septopetal hypothalamic substance P cells do not contain GABA or acetylcholine; (iii) substance P fibers from pericellular baskets around calbindin-containing lateral septal neurons with a high degree of selectivity; (iv) approximately 90% of the entire calbindin cell population are postsynaptic targets of substance P axons; (v) their terminals contact the soma and the dendrites of these cells, among them the somatospiny neurons; and (vi) the extrinsic substance P boutons establish asymmetric, while the intrinsic substance P axon terminals form symmetric membrane specializations. Because neurons in the lateral septal area receive hippocampal input and project massively to hypothalamic areas, the different types of substance P input on these neurons can modify the information flow arriving from the hippocampus to diencephalic brain structures at the level of the lateral septal area.

Hypothalamic Leu-enkephalin-immunoreactive fibers terminate on calbindin-containing somatospiny cells in the lateral septal area of the rat.

Correlated light and electron microscopic double-immunostaining experiments for Leu-enkephalin and calbindin were employed to determine the postsynaptic targets in the septal complex of Leu-enkephalin fibers. Chronic surgical isolation of the septal complex from its hypothalamic afferents and retrograde tracer studies using wheat germ agglutinin-conjugated horseradish peroxidase, both followed by an immunostaining for Leu-enkephalin, were performed to elucidate the location of the origin of these axon terminals. Furthermore, a colocalization study for glutamic acid decarboxylase and Leu-enkephalin was carried out on hypothalamic sections to determine their possible coexistence in cells projecting to the lateral septum. These studies revealed that 1) Leu-enkephalin-immunoreactive axons form pericellular baskets around a population of lateral septal area neurons; 2) they establish exclusively asymmetric synaptic contacts on their soma and initial dendritic segments; 3) 10% of the lateral septal area calbindin-containing cells, which are all of the gamma-aminobutyric acid (GABA)-ergic somatospiny type, are innervated by Leu-enkephalin-immunoreactive baskets; 4) only 40% of the Leu-enkephalin target neurons are calbindin immunopositive; 5) the septopetal Leu-enkephalin fibers derive from neurons located in the ipsilateral perifornical area and anterior hypothalamus; and 6) none of their cells of origin cocontains the inhibitory transmitter GABA. These observations indicate that hypothalamic Leu-enkephalin-containing neurons are non-GABAergic excitatory cells. Hence, they can effectively stimulate a population of lateral septal area neurons, including the somatospiny cells, which are all GABAergic. Therefore, after stimulatory Leu-enkephalin action, these neurons can inhibit their postsynaptic targets, including other projective lateral septal neurons.

Neuropeptide-Y innervation of estrogen-induced progesterone receptor-containing dopamine cells in the monkey hypothalamus: a triple labeling light and electron microscopic study.

Light and electron microscopic triple immunostaining was performed on coronal vibratome sections prepared from the hypothalamus of ovariectomized (OVX) and OVX plus estrogen-treated African green monkeys (Cercopithecus aethiops). Immunoreactivity for progesterone receptors (PRs) and neuropeptide-Y (NPY) was visualized by a dark blue to black nickel diaminobenzidine reaction, while the tyrosine hydroxylase-containing perikarya were labeled with a light brown diaminobenzidine reaction. In the OVX plus estrogen-treated material, 30% of the tyrosine hydroxylase-immunoreactive neurons contained PR-immunopositive nuclei. The majority of these cells were found in the central portion of the periventricular area, and a few could be observed in the anterior hypothalamus and the arcuate and dorsomedial hypothalamic nuclei. These tyrosine hydroxylase-immunoreactive PR-containing cells were surrounded with NPY-immunoreactive axon terminals. A correlated electron microscopic analysis of the same sections revealed synaptic contacts between these NPY-immunoreactive boutons and the PR-containing tyrosine hydroxylase-immunoreactive neurons. In contrast, in the OVX animals, no PR-containing tyrosine hydroxylase-immunoreactive neurons could be detected. In these monkeys, the frequency of synaptic contacts between the NPY-immunoreactive axon terminals and tyrosine hydroxylase-immunopositive cells was similar to that in the OVX plus estrogen-treated monkeys. These observations indicate that in a population of hypothalamic dopamine cells, the presence of nuclear PRs is estrogen dependent, show that these cells are innervated by NPY axons, and suggest that these estrogen-induced PR-containing dopamine neurons are involved in mediation of the effect of NPY on hypophyseal hormone secretion, including ovarian steroid hormone-dependent LH and PRL release.

Estrogen induces ultrastructural changes in progesterone receptor-containing GABA neurons of the primate hypothalamus.

Estrogen affects gonadotrophin levels and sex behavior in monkeys. This action could be via inhibitory GABA-ergic neurons in the hypothalamus. We tested for direct estrogen actions on such neurons. Seven days after ovariectomy (OVX) or OVX + estrogen treatment (10 mg estradiol valerate in 1 ml sesame oil s.c. on the day of OVX), light- and electron-microscopic double immunostaining procedures were used for simultaneous visualization of immunoreactivity for progesterone receptors (PR) and glutamic acid decarboxylase (GAD), and to detect ultrastructural changes in PR-containing neurons in the arcuate and ventromedial hypothalamic nuclei of colchicine- and noncolchicine-treated African green monkeys (Cercopithecus aethiops). Immunoreactivity for PR was found only in cell nuclei, and estrogen treatment enhanced the intensity of the immunostaining: in estrogen-treated monkeys in the arcuate nucleus 62%, while in the ventromedial nucleus 42% of the neurons contained PR-immunoreactive nuclei. All of the PR-containing neurons were immunopositive for GAD in colchicine-pretreated monkeys. OVX induced whorl body formation, while estrogen treatment of OVX animals resulted in a large number of nematosomes. While all of the whorl bodies and the majority of nematosomes were observed in PR-immunopositive GAD neurons, nematosomes were also found in non-PR-containing GAD-immunoreactive cells.